Blood Levels of Liver Cancer Drug Lenvatinib Predict Who Responds to Treatment

Lenvatinib has become one of the most important weapons against hepatocellular carcinoma, the most common form of liver cancer and one of the deadliest malignancies worldwide. As a first-line therapy for patients whose tumors cannot be removed surgically, the oral kinase inhibitor has extended survival for many. Yet clinicians have long wrestled with a frustrating reality: some patients respond dramatically while others derive little benefit, and the reasons have remained murky. A new multi-omics study published in the Journal of Translational Medicine now offers a compelling explanation, tracing treatment failure back to a surprisingly simple variable—how much of the drug actually circulates in a patient’s bloodstream—and connecting that variable to sweeping changes in tumor metabolism and immune architecture.

The research team, led by investigators at Sichuan Cancer Hospital and the University of Electronic Science and Technology of China, prospectively enrolled 194 patients with hepatocellular carcinoma who were receiving lenvatinib. Rather than treating the drug as a one-size-fits-all prescription, the researchers measured plasma drug concentrations and correlated them with clinical outcomes. Their analysis converged on a candidate threshold: a trough concentration—the lowest drug level between doses—of 26.5 nanograms per milliliter. Patients whose trough levels fell below this line were markedly less likely to achieve an objective response, while those above it fared substantially better.

What elevates this finding from a routine pharmacokinetic observation is the predictive power of the resulting model. The investigators combined the 26.5 ng/mL concentration threshold with tumor size to build a composite predictor, and it performed impressively across independent patient groups. In the training cohort the model achieved an area under the curve of 0.884, and it held up in internal validation at 0.868 and in an external cohort at 0.873. Values in this range suggest a tool that could meaningfully guide clinical decisions, flagging early on which patients may need dose adjustments or alternative strategies rather than waiting months to discover that a regimen is failing.

The team also searched for routine laboratory markers that might serve as accessible proxies for drug exposure. Among the clinical variables examined, only uric acid showed a significant correlation with lenvatinib levels, with a correlation coefficient of 0.657 and a p-value below 0.001. While uric acid is hardly a perfect surrogate, the strength of this association is notable and hints that common metabolic markers could eventually help clinicians estimate drug exposure without specialized assays, at least as a preliminary screen in resource-limited settings.

To understand why sub-therapeutic exposure undermines treatment, the researchers turned to untargeted plasma metabolomics, cataloging the small molecules circulating in patients’ blood. The results revealed that inadequate lenvatinib exposure was accompanied by a distinct form of lipid metabolic reprogramming. Most striking was the accumulation of acylcarnitines, molecules that arise when fatty acids are shuttled into mitochondria for oxidation. Acylcarnitine buildup is a well-recognized signature of incomplete or dysregulated fatty acid metabolism, and its presence in patients with low drug exposure suggests that insufficient lenvatinib levels allow tumor-associated lipid handling to shift into a state that favors survival and progression rather than metabolic stress and cell death.

The metabolic story was reinforced at the cellular level. The team validated their metabolomic findings in vitro using cell-based experiments, and they also sequenced microRNAs carried by exosomes—tiny membrane-bound vesicles that cells release into circulation as a form of intercellular messaging. Together, these approaches painted a coherent picture in which lipid metabolism is not merely a bystander in lenvatinib resistance but an active participant, rewiring how tumor cells and their supporting environment generate energy and building blocks under the pressure of inadequate drug exposure.

Perhaps the most visually striking discovery came from spatial analysis of the tumor microenvironment. Using multiplex immunohistochemistry, which allows several immune markers to be visualized simultaneously on tissue sections, the researchers found that sub-therapeutic drug exposure produced what they describe as a hot but suppressed tumor microenvironment. Tumors in this state were densely infiltrated with CD8-positive T cells, the cytotoxic soldiers of the immune system, and showed high densities of PD-L1-positive cells, the molecular brake that tumors use to disable immune attack. On paper, such a tumor should be inflamed and vulnerable. In practice, it was immunologically paralyzed.

The reason for that paralysis lay in the spatial arrangement of the cells. The multiplex imaging revealed extensive colocalization between CD8-positive T cells and CD163-positive macrophages, a population of tumor-associated macrophages generally associated with immunosuppression. Rather than engaging tumor cells, the T cells appeared to be physically entrapped by these macrophages, held in place and functionally restrained. This is a spatial form of immune exclusion: the anti-tumor cells are present in abundance but are corralled by suppressive myeloid cells, preventing them from mounting an effective attack. The finding reframes resistance not as an absence of immune activity but as a misdirection of it.

Single-cell RNA sequencing added a systems-level dimension to the story. By profiling gene expression in individual cells, the researchers identified CD8-positive T cells as the primary hub of intercellular communication within the tumor ecosystem, linking the metabolic shifts detected in plasma to the spatial patterns of immune activation observed in tissue. This positions the exhausted, macrophage-entangled T cell as the central node where metabolism and immunity converge—a molecular meeting point through which inadequate drug exposure translates into clinical resistance.

The implications for practice are direct. The authors advocate for therapeutic drug monitoring of lenvatinib, using trough concentration measurements to keep patients above the 26.5 ng/mL threshold, and they point toward combination strategies that target the metabolic-immune axis uncovered by their analysis. If sub-therapeutic exposure drives acylcarnitine accumulation and macrophage-mediated T-cell entrapment, then pairing dose optimization with agents that reprogram lipid metabolism or unleash macrophage-suppressed T cells could convert non-responders into responders. For a cancer that remains among the leading causes of cancer death globally, a measurable, modifiable predictor of response—backed by a mechanistic account of why it matters—represents exactly the kind of translational progress that turns a standard therapy into a reliably effective one.

Subject of Research: Lenvatinib drug exposure, lipid metabolic reprogramming, and immune microenvironment remodeling in hepatocellular carcinoma treatment response

Article Title: Lenvatinib exposure is associated with therapeutic response in HCC, correlating with lipid metabolic reprogramming and spatial remodeling of the immune microenvironment

Article References: Wang, H., Chen, Y., Liu, A., Zhou, Y., Yang, X., Chen, P., Jiang, Q., Zhang, J., Bo, W., Hou, G., Liu, J., Zhang, H., Liu, H., Xiao, H., Jiang, S., Feng, X., Xu, B., & Chen, Y. (2026). Lenvatinib exposure is associated with therapeutic response in HCC, correlating with lipid metabolic reprogramming and spatial remodeling of the immune microenvironment. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-09056-3

Image Credits: AI Generated

DOI: 10.1186/s12967-026-09056-3

Keywords: hepatocellular carcinoma, lenvatinib, therapeutic drug monitoring, pharmacokinetics, lipid metabolism, acylcarnitines, tumor microenvironment, CD8 T cells, tumor-associated macrophages, PD-L1, single-cell RNA sequencing, drug resistance